Premium
Does the Index of Ideality of Correlation Detect the Better Model Correctly?
Author(s) -
Toropova Alla P.,
Toropov Andrey A.
Publication year - 2019
Publication title -
molecular informatics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.481
H-Index - 68
eISSN - 1868-1751
pISSN - 1868-1743
DOI - 10.1002/minf.201800157
Subject(s) - quantitative structure–activity relationship , minnow , pimephales promelas , molecular descriptor , correlation , data mining , set (abstract data type) , computer science , software , machine learning , artificial intelligence , mathematics , fish <actinopterygii> , biology , geometry , fishery , programming language
The CORAL software is a tool to build up predictive models for various endpoints by means of Quantitative Structure‐Property/Activity Relationships (QSPRs/QSARs). A new criterion for assessment of the predictive potential of QSPR/QSAR models, so‐called Index of Ideality of Correlation ( IIC ) is applied to improve the software. The ability of the IIC to detect models with better predictive potential is checked up with groups of random splits of data into the structured training set and extrenal validation set. To this end, two endpoints are examined (i) Toxicity towards Fathead minnow ( Pimephales promelas ) ; and (ii) drug load capasity of samples “micelle‐polymer”. Applications of the IIC for endpoint represented by traditional Simplified Molecular Input‐Line Entry System (SMILES) together with so‐called quasi‐SMILES has shown the suitability of the IIC be a tool to detect better model.